Background Genetic markers of susceptibility for anxiety and depression have been elusive. Increased sample size, utilizing symptoms rather than disorders and, exploring genetic covariability between pairs of traits are known approaches to increase the power of GWAS and have been used isolatedly. We have previously demonstrated a strong genetic covariability between depression and anxiety symptoms in a family based cohort, The Baependi Heart Project. Combining two strategies, symptoms approach and genetic covariability in a GWAS can result in greater detection power of genetic markers associated with those traits. Methods We investigated depression and anxiety symptoms, quantified using the Hospital and Depression Scale (HADS) in 1,375 individuals from 93 nuclear families recruited from an admixed population in Brazil. GWAS was performed for anxiety and depression scores both independently (univariate analysis), and using bivariate analysis taking into consideration the genotypic covariance between those traits (rhog=0.81). The Baependi cohort was genotyped using different GWAS platforms (custom arrays to capture the tri-ancestry genetic structure of the Brazilian population; Affymetrix Axiom Incor array and Affymetrix SNP chip 6.0; Affymetrix, Santa Clara, CA). Results The univariate approach revealed interesting genetic targets, but none with genome-wide significance. However, the bivariate approach identified significant hits (p Discussion Taking into consideration the genetic co-variability between anxiety and depression symptoms added power for the identification of genetic markers associated with these traits.